By Gloria Ogbonna
The Trump administration has taken a major step to tighten control over cutting-edge artificial intelligence technology developed in the United States, moving to restrict foreign access to some of Anthropic’s most advanced AI systems amid growing national security concerns.
According to a report published Friday by Axios, the administration has ordered new export controls on Anthropic’s flagship AI models, Mythos 5 and Fable 5, which are considered among the most sophisticated artificial intelligence systems currently available. The restrictions are intended to prevent foreign governments, companies, and individuals from accessing the technology without government approval.
The report states that Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei on Friday formally notifying the company that both AI models would be placed under federal export-control regulations.
The move marks one of the most significant actions yet by the U.S. government to regulate the international distribution of advanced artificial intelligence technology, reflecting increasing concerns in Washington that powerful AI systems could be exploited by foreign adversaries for military, intelligence, cybersecurity, or economic purposes.
Under the reported restrictions, Anthropic must obtain federal licenses before exporting, re-exporting, or transferring the affected AI models. The controls reportedly apply not only to users located outside the United States but also to foreign nationals residing within the country.
Sources familiar with the matter told Axios that administration officials had previously requested that Anthropic delay the public release of the models while security concerns were reviewed. However, the company proceeded with the launch earlier this month, prompting federal officials to impose export restrictions after the systems became publicly available.
Government officials reportedly became increasingly concerned after another company demonstrated what was described as a successful “jailbreak” of Mythos 5. A jailbreak occurs when users discover methods to bypass built-in safety protections and limitations within an artificial intelligence model, potentially allowing it to generate prohibited or dangerous outputs.
According to administration officials, the discovery raised questions about whether foreign actors could exploit vulnerabilities in the system for purposes that might threaten U.S. national security interests.
The Commerce Department’s action is part of a broader effort by the administration to protect strategically important technologies from falling into the hands of rival nations. In recent years, Washington has expanded restrictions on advanced semiconductor exports, high-performance computing systems, and AI-related hardware as competition with China and other foreign powers intensifies.
Anthropic confirmed Friday that it had received the government directive and announced that it would immediately suspend access to both Mythos 5 and Fable 5 while complying with federal requirements.
Despite agreeing to follow the order, the company publicly challenged the government’s justification for the restrictions.
“We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users,” Anthropic said in a statement.
The company argued that the reported jailbreak involved only limited vulnerabilities and did not warrant pulling back a commercial product already being used by millions of people around the world.
“However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” the statement continued.
Anthropic further warned that if similar standards were applied across the AI industry, it could significantly slow innovation and prevent companies from releasing new frontier AI systems.
“If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” the company said.
The AI developer also emphasized the extensive safety testing conducted before the release of the models. According to Anthropic, thousands of hours were spent evaluating the systems in collaboration with government partners in both the United States and the United Kingdom, as well as third-party researchers and the company’s own internal safety teams.
The company said those evaluations failed to uncover any universal jailbreak capable of broadly bypassing the models’ safeguards. Instead, Anthropic characterized the identified vulnerabilities as limited in scope, largely harmless, or not unique to Mythos 5.
Mythos 5 and Fable 5 were officially launched on June 9 and represent the latest generation of Anthropic’s AI technology. The models are designed to perform highly complex tasks, including software development, scientific analysis, advanced reasoning, cybersecurity operations, and research assistance.
Anthropic has described Mythos 5 as possessing some of the strongest cybersecurity capabilities available in any commercial AI model, while Fable 5 was developed as a more heavily safeguarded version intended for broader deployment.
The administration’s decision comes amid increasing debate over the rapid advancement of artificial intelligence and its potential impact on national security.
Earlier this year, reports indicated that Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an emergency meeting with leading Wall Street executives to discuss cybersecurity concerns linked to advanced AI systems, including Anthropic’s Mythos platform. Officials reportedly sought to better understand how increasingly capable AI models could affect financial markets, critical infrastructure, and digital security.
Experts have warned that as artificial intelligence becomes more powerful, governments may face growing pressure to balance innovation with security concerns. Supporters of tighter controls argue that advanced AI systems could provide strategic advantages to hostile nations if left unrestricted, while critics contend that excessive regulation could hinder technological progress and weaken America’s competitive edge in the global AI race.
As of Friday evening, the Commerce Department had not publicly commented on the reported restrictions or responded to requests for additional information.
The new export controls highlight the growing role artificial intelligence is playing in global geopolitical competition and signal that Washington may continue expanding oversight of advanced AI technologies as their capabilities—and potential risks—continue to grow.
Source Breitbart
Original article
Comments